Roman Numerals Converter

Reading is forgiving, writing is strict

Look at a traditional clock face and the four is almost always IIII, not IV. That is not a mistake, and it is not rare — it is on Big Ben, on most tower clocks, and in a great deal of medieval and early modern writing.

So a converter has two competing obligations. Refusing IIII would be wrong about history: the numeral exists, it is unambiguous, and someone typing it has almost certainly copied it off something real. But emitting IIII would be wrong about convention, because standard modern form is IV and that is what a reader now expects.

This tool does both. It reads IIII, VIIII and the other additive spellings and tells you the value, then offers the standard form as a note rather than an error. It only ever writes standard form. And it stays quiet when the input is already standard, because a note on every numeral would be a note on none.

Why clocks kept IIII is not settled. The most durable explanation is visual balance: with IIII at four, the left side of the dial uses only I, the bottom only I and X, and the whole face is more symmetrical than it is with IV. The story about Louis XIV preferring it is repeated often and evidenced poorly.

Why there is no Roman zero

There is no Roman numeral for zero, and the tool refuses the input rather than returning something. That is worth explaining rather than just enforcing, because the absence is not an oversight.

Roman numerals are an additive tally system, not a positional one. Each symbol carries its own fixed value wherever it sits, and the number is what the symbols add up to. A positional system — the one we use — needs a zero as a placeholder, because 105 and 15 differ only by an empty column. An additive system has no columns to leave empty, so it never needs the symbol.

Zero as a number in its own right arrived in Europe with Hindu-Arabic numerals, by way of Indian mathematics and Arabic scholarship, and it arrived together with the positional notation that requires it. Medieval scholars who needed to write nothing in a Roman context used the word nulla instead.

The same reasoning rules out negatives and fractions: the notation has no sign and no fractional part. Romans did handle fractions, but in a separate duodecimal system of twelfths with its own names — uncia, semis — which is not this notation and is not offered here.

The subtractive pairs

There are exactly six places where a smaller symbol sits before a larger one and gets subtracted: IV (4), IX (9), XL (40), XC (90), CD (400) and CM (900). Everything else is additive.

They exist to enforce another rule: no symbol repeats more than three times. Without subtraction, four would be IIII and nine VIIII, and ninety would be LXXXX. The subtractive pairs keep every numeral to at most three of anything — which is why the two rules are really one rule, and why a numeral containing four identical symbols in a row is a signal you are looking at an older or informal spelling.

A second rule follows from the same logic: V, L and D never repeat at all. Two of any of them would simply be the next symbol up — VV is X, LL is C, DD is M — so a standard numeral never contains more than one of each. Both rules are asserted across all 3,999 values in this page’s test suite.

The subtraction is also restricted in which pairs are allowed. Only I before V or X, X before L or C, and C before D or M. IL for 49 and IC for 99 look plausible and are not standard; 49 is XLIX and 99 is XCIX.

Where the numbering stops

Standard notation runs out at 3,999 — MMMCMXCIX. The reason is the three-repeat rule meeting the largest symbol: M is 1,000 and there is nothing above it, so three Ms is as far as thousands go.

Romans did write larger numbers, using a vinculum— an overbar over a numeral meaning “multiply by a thousand”, so V̄ is five thousand and X̄ is ten thousand. That is a typographic mark rather than a letter, and it cannot be written in plain text; substitutes using brackets or parentheses exist but are not standard and are not universally read the same way.

So this tool refuses anything above 3,999 and says why, rather than producing a bracket notation that would look authoritative and might not be understood by whoever reads it next. If you need a larger value, the honest answer is that Roman numerals are the wrong notation for it.

In practice the ceiling is rarely a problem, because the places numerals are still used are all under it: years on cornerstones and film credits, monarch and papal regnal numbers, Super Bowl editions, book chapter and preface pagination, and clock faces.

Doing it in a spreadsheet

Excel and Google Sheets have =ROMAN(A1), which produces standard form for 1 to 3,999 and errors outside it — the same boundary, for the same reason. It takes an optional second argument for progressively more concise non-standard forms, which is best left alone unless you specifically want them.

Going back, Excel has =ARABIC(A1). Google Sheets does not, so the practical route there is a lookup table or a short script — and a lookup table is usually enough, since a sheet dealing with numerals is normally dealing with years or chapter numbers rather than arbitrary values.

One caution if you are storing numerals as data: they sort alphabetically, not numerically, so a column of years in numerals will sort IX before V. Keep the number in its own column and treat the numeral as presentation, exactly as you would with any other formatted value.

Related calculators

Other ways of writing the same quantity:

More in Conversion, or browse all calculators.

Sources and methodology

Nothing here is fetched and there is no data feed. There are also no conversion factors — this is a transcription between two notations for the same integer, so what the sources settle is the convention: the seven symbols, the subtractive pairs, the rule that no symbol repeats more than three times, and the overbar notation that standard plain text cannot express.

• Roman numerals — the subtractive principle and standard modern formEncyclopaedia Britannica · verified 2026-09-06 · The seven symbols, the subtractive pairs, and the convention that no symbol repeats more than three times
• ISO 8601 and the case for unambiguous date notationInternational Organization for Standardization · verified 2026-09-06 · Why numerals used decoratively on dates — film copyright lines, cornerstones — are not a machine-readable date format
• The vinculum and apostrophus notations for large valuesEncyclopaedia Britannica · verified 2026-09-06 · That values above 3,999 required an overbar multiplying by a thousand, which is why standard plain-text notation stops there